bmad-party-mode

Orchestrate BMAD subagents in parallel for roundtable discussions.

Updated May 3, 2026
One-click install
npx skills add https://github.com/tabesink/deepdoc-agent --skill bmad-party-mode-tabesink
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: bmad-party-mode
Source: https://github.com/tabesink/deepdoc-agent/tree/main/.cursor/skills/bmad-party-mode
Command: npx skills add https://github.com/tabesink/deepdoc-agent --skill bmad-party-mode-tabesink

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill turns a single-agent chat into a genuinely diverse roundtable by orchestrating multiple BMAD agents as independent subagents, reducing performative similarity and surfacing disagreements and complementary expertise.

Core Features & Use Cases

  • Independent multi-agent participation: Spawns 2–4 selected BMAD agents as real subagents (not roleplayed characters), each producing its own unabridged response.
  • Config-driven agent roster & context: Loads greeting/language from core config and builds a voice roster from the agent manifest, optionally injecting project-context when available.
  • Round-based orchestration loop: Picks voices per user message, runs a parallel spawn, and presents distinct agent sections while keeping shared “discussion context” compact.

Quick Start

Activate party mode by telling the assistant to run bmad-party-mode for a roundtable discussion about your topic, optionally adding --solo if you want responses generated without subagents.

Frequently Asked Questions about bmad-party-mode

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a multi-agent roundtable discussion with different perspectives?

A multi-agent roundtable discussion is run by orchestrating independent BMAD subagents in parallel, each generating an unabridged response in a distinct section. The Skill selects appropriate agents per message from an agent manifest to ensure diverse viewpoints.

What is parallel multi-agent orchestration and how does it facilitate cross-talk?

Parallel multi-agent orchestration spawns 2–4 independent subagents simultaneously to generate diverse perspectives on a topic. It facilitates cross-talk and reactions by presenting each agent's complete response in its own section while maintaining compact shared discussion context.

Can I generate multiple agent viewpoints without spawning real subagents?

You can generate multiple agent viewpoints without subagents by using the solo mode. Adding the solo flag instructs the system to roleplay all selected BMAD agents within a single response cycle instead of spawning independent parallel subagents.

Do I need an agent manifest to start a collaborative discussion?

You need an agent manifest and core config to start a collaborative discussion. The Skill loads greeting and language settings from the core config and builds the voice roster of available agents from the manifest, optionally injecting project context when available.

What's the best way to reduce performative similarity in AI agent conversations?

The best way to reduce performative similarity is to orchestrate multiple BMAD agents as independent subagents rather than roleplaying characters. This approach surfaces disagreements and complementary expertise by producing genuinely diverse, unabridged responses from separate execution contexts.

How many agents can participate in a single roundtable orchestration loop?

A single roundtable orchestration loop spawns 2–4 selected BMAD agents as real subagents per user message. The loop picks voices, runs the parallel spawn, and presents distinct agent sections while keeping the shared discussion context compact across successive rounds.